I'm glad to read that Google has confirmed what I have been preaching "Panda flavored main ranking algorithm" since January 2012 on both threads: Google Updates and SERP Changes - Jan 2012 & Google Updates and SERP Changes - Feb 2012 :
|I think there hasn't been a Panda update in January 2012 and Google hasn't announced any Panda update for sometime. |
However I think that the factors of Panda are already integrated in the current main ranking algorithms "Panda flavored main ranking algorithm".
|making it more integrated into our pipelines |
So this sounds like the Panda Dance maybe coming to an end. I didn't think Google could live with that too long - it was so 2002!
Sounds like we should expect to notice fewer Panda data refreshes in future? Are they (not) saying these will be folded more frequently and incrementally into the data to avoid the 'dance' phenomenon?
|For me the big one was something we've been conjecturing about here -some kind of new integration between Panda and the rest of the algorithm |
From what i see we're not seeing any increase in the intervals between iterations, which still seems to command the overiding "quality score".
Something a few others and myself noticed in early January was the ability for a new quality page to rank quickly on a Pandalized site , but i think this is within the existing "quality score" - so that page and site may still be feeling the effects of the Panda penalty until a full iteration is carried out
It would be great to see the quality score updated on the fly, but i just don't see that for a while, if ever. Thoughts?
As I said, having periodic and infrequent data refreshes/ metric recalculations is not where Google's been headed with all it's infrastructure improvements over last 2-3 years. Can't last, can it?
Whitey, toward the end of January there were also reports of long time Pandalized pages showing some movements in the rankings on several different dates.
Tedster, in the context of this , what do you make of Gyppo's 6 week start to finish revival example [ 8th Dec to 19 Jan ), excluding minor earlier works ? - [webmasterworld.com...]
- looks to me like something was speeded up and in play per Reseller's claims and your observation with various other members and threads. I'd just like to get a sense of what elements were integrated into that "pipeline". Gyppo still relied on the data refresh ultimately.
There's not enough data to say much that's truly definitive, Whitey. On the surface, the few reports we've got seem to say individual URLs are allowed to move in the rankings a bit more, but the domain as a whole "might" still carry its overall Panda ranking.
But don't run off to the bank with that idea just yet. It's way too small a sampling and what we do have is anecdotal. We can't even be sure each of these reporting sites were actually suffering from Panda.
I've felt that there's been an ongoing correspondence between the capabilities of the search interface and the capabilities of the algorithm, as they both pull from improvements in the underlying index. As I read this list, it jumps out at me that the algorithm is most likely becoming more granular, as an AI algorithm would become over time, and also faster in making subtle distinctions.
And assuming (as I do) that Google is triangulating everything ;) , almost all of 17 announced quality highlights suggest increased granularity... but here are some I'd single out...
|Faster autocomplete. ...improvements to our autocomplete system to deliver your predicted queries much faster. |
|Autocomplete spelling corrections. ...This launch targets corrections where the spelling change is very small. |
|Better spelling corrections for rare queries. ...improves one of the models that we use to make spelling corrections... for a number of rare queries. |
|Improve detection of recurrent event pages. ...improvements to how we determine the date of a document.... particularly for pages discussing recurring events. |
|High-quality sites algorithm improvements. ...improved how Panda interacts with our indexing and ranking systems, making it more integrated into our pipelines. |
|More relevant related searches. ...we’ve updated the model for generating related searches, resulting in more useful query refinements. |
I can see it becoming a real-time "fluctuations" sort of job now. So where they were releasing new panda versions every month or two they will make mini changes almost daily,
|Whitey. On the surface, the few reports we've got seem to say individual URLs are allowed to move in the rankings a bit more, but the domain as a whole "might" still carry its overall Panda ranking. |
That's what I'm seeing too - recovered some rankings, include top level ones over the past week, but traffic still 30% of what it was. Although I suspect changes to long tail searches could be an answer to that.